Scatter search techniques applied to electromagnetic problems

This paper presents a new multiobjective scatter search algorithm. First, one demonstrates the standard scatter search performance by applying it to solve a single-objective inverse electromagnetic scattering problem. Then, the changes necessary to solve multiobjective optimization problems that should be implemented in a single-objective algorithm are discussed. As a result of this discussion, one proposes the use of a nondominated sorting technique and a niched-type penalty method in order to extend the scatter search technique to multiobjective problems. The results obtained by the new algorithm show its performance and helpfulness in solving multiobjective electromagnetic problems.